Speech-based user interfaces are employed in a variety of applications. In such user interfaces, the system's output to the user is spoken and audible, rather than displayed. User input to these systems may also be spoken, recognized by automatic speech recognition (ASR) technologies, or by key input, most commonly dual-tone multi-frequency (DTMF) keys on telephones.
One of the most common and widely used speech-based user interface is the telephone-based access to voice messages that millions of persons use every day. In such a system, a user calls into a voice mail system, gains access to the system, and hears a status message describing the current state of messages in the mailbox. (This status message will be referred to hereinafter by the term “summary”.) The user is then offered a fixed set of functions. FIG. 1 shows typical steps involved in such an interaction in prior art systems.
FIG. 1 illustrates the steps involved in the use of a prior art speech-based interface, such as that used in a voice mail system. In accessing step 110, a user enters the system by dialing an access number. In authentication step 120, the system performs user authentication by prompting the user for a user id and password combination, or some other login procedure. Following successful authentication, step 130 consists of presenting a fixed, pre-programmed summary to the user that offers information regarding the status of the messages in the user's mailbox. Next, in step 140, a fixed menu of system functions that are available to the user is presented. This menu may contain nested menus, each offering groups of functions presented hierarchically. The user may then, in step 150, access their choice of system functions offered in the fixed menu.
Other prior art speech-based systems are intended primarily for use in a hands free/eyes free environment, such as an automobile. In such as system, a user would activate the car, hear a status message, and have spoken control of automobile functions, such as radio, climate control, navigation systems, etc. via the speech-based system.
Both these messaging and automobile prior art systems have a major element in common. Although both were formerly standalone, isolated systems, current implementations of such systems are now in communication with the rest of the world via the internet.
The messaging system has evolved from simple voice messages, to unified messaging (including e-mail and facsimile, and soon to include video), to unified communications, where the user not only accesses their messages, but initiates calls, manages their schedule, and has access to a voice browser for the world wide web.
The automobile system, while less mature in its development, is already evolving, with functionality being available that permits a driver to access web-based information without taking their hands off the steering wheel.
Both of these exemplary prior art systems have another element in common. Both of the user interfaces present summary information and offer functions, each of which is programmed in advance by the manufacturer. The summary information is based on a standard template. The functions available within a standard voice menu are based on a pre-set hierarchy or voice command vocabulary structure.
As the wealth of potential information and available functions increases, some users'needs cease to be appropriately addressed by these prior art solutions. For instance, the pre-defined summary may include information the user does not desire, and may not include other information that is important to the user. Also, each user tends to use only a subset of these numerous available functions. Nevertheless, each user is forced to unnecessarily navigate amongst the “clutter” to access those functions that are used most often. Furthermore, there is no ability to create new functions, such as those that would retrieve specific elements of information from the web that are of interest to the particular user.
While customization exists in other domains (e.g. web services and PC-based GUI applications), it has long been held that customization was impractical for speech interfaces, given their very limited nature. However, in view of recent expanded uses of speech-based interfaces, it would be desirable to have a system and method that would allow for user customization of these interfaces, thereby facilitating a user's experience with the systems that they control.
One particular prior art system exists that attempts to address some of these drawbacks. On Apr. 10, 2000, in a company press release, Tellme Networks Inc. of Mountainview, Calif. announced a telephone-accessed service named Tellme™ which, at that time, was accessible via a toll-free number and customizable either by telephone or on the www.tellme.com web site. The Tellme service allows a user to reorder a set of preset elements in an information summary, and to select the specific data items that are used to quantify those elements. This system only permits a user to select a function from a plurality of available functions. The system does not permit the user to define a new function nor does it permit the user to define a modified version of an existing function. Because of this limitation, the system does not permit the user to modify the contents or presentation of the information summary that is played upon accessing the system. Furthermore, the user cannot customize the interface itself. The present invention seeks to overcome the drawbacks of prior art systems.